An analysis of car chassis
Contents
1.0Introduction
2.0General results
3.0The Lowcost Chassis
4.0Larger engines in the Lowcost chassis
5.0Regarding ladder frames
6.0A list of chassis analysis and quoted results
7.0A design checklist for space frames and ladder frames
8.0Regarding monocoques
9.0Lowcost book errors
10.0Suspension Geometry
11.0Analysis techniques
1.0Introduction
My analysis of car chassis structures began as part of my interest in building my own car. There are very few books available that explain the design issues facing the amateur chassis builder. Most available books are very brief in this area and concentrate more on other aspects of car design and construction such as suspension and bodywork. The few books that do deal with chassis design do not always make it clear how the advice given can be applied to a real design that an amateur builder could realistically attempt to make. Even fewer books give any indication of what torsional stiffness values are likely to be achieved.
Major car manufacturers make use of a form of analysis called Finite Element Analysis to help them design their chassis. I have several years experience of designing and testing light weight structures and using finite element analysis to help predict and understand test results. I also have more general experience of this type of analysis and have been involved in various projects were it was applied. I decided to use it to gain some understanding of the various issues of chassis design relating to what an amateur builder could realistically attempt.
My analysis is based on simple models that are intended to represent the main load bearing components of a chassis. Minor brackets, the effect of fitting the drive train, and other parts or regions of the car that make little difference to the basic chassis structure are not included. Panels attached by rivets are not included as rivets can work loose over time and lose their structural capability even though they may still be perfectly adequate to support non structural panelling. The results of my analysis will consequently be different to a real chassis. The quoted results for stiffness and weight should still be reasonably close to reality and adequate for showing the effect of different designs at a basic level.
Please note that the information given in chapter 9 is copied from the Internet and that I have not fully checked it.
While the statements made in this document are produced and provided in good faith I accept no liability whatsoever for any damage, injury or loss resulting from their use or from any incorrect or unintended interpretation of the information.
2.0General results
The basic way of assessing a chassis design is to establish its torsional stiffness. The torsional stiffness is stated as the torque required to twist the chassis by a given amount. One of the most common styles of space frame is the Lotus or Caterham Seven type. Most kitcars based around Seven type chassis are in the region of 1000 to 2000 ftlbs per degree of twist. Most small mid engined car space frame designs of similar complexity also fall in this range.
A lot of builders believe that extra bracing in the form of diagonals or welded in panels will just add weight and simply omit it. This can seriously reduce the stiffess of a spaceframe chassis to the point where even a very basic ladder frame would easily beat it. A properly braced frame can be made out of smaller or thinner walled tubes for the same stiffness thus more than compensating for the extra tubes. My modifications for the Lowcost actually reduce the overall number of tubes and the total weight.
My initial analysis showed that a lot of space frame chassis were less effective, weight for weight, than properly designed X braced ladder frames. The reason for this is that most space frames are far from optimum and that a simple X braced ladder frame is better than most people think. This may surprise you as a lot of hype surrounds space frames and ladder frames seem to have a poor image. This situation is not really justified and it is worth noting that the main structural members of the Lotus Elise chassis are two large beams that are connected by other beams and panels to make a structure that could be regarded as a very sophisticated ladder chassis or as a hybrid ladder frame and monocoque chassis.
Overall a properly designed simple space frame has a small advantage, probably about five percent or less, in weight and stiffness for the complete car over an equally well designed simple X braced ladder frame. A more complex space frame with fully triangulated front and rear suspension regions, engine bay and sill structures may give more advantage but will be harder and more expensive to make. This difference gives an advantage in race cars especially when the budget and analysis skills to produce a good design are available. However for most road cars and some race cars a ladder frame would be perfectly adequate.
Space frames have an additional advantage in both weight and complexity where the chassis panelling forms a significant portion of the bodywork as on many Seven type cars. This maximises the advantages of a space frame by removing the need for the extra weight and complication of additional bodywork and the extra structures needed to support it.
Advantages of the ladder frame that are often overlooked are that the available space and ease of access to mechanical parts is often better and engine exhaust systems are less likely to be restricted by the need to route them around chassis tubes. Additional structures are often required with ladder frames to support bodywork but these can often be designed to brace the basic chassis structure.
The poor image of ladder frames may be due to early chassis design practice which was to use C section chassis rails instead of fully boxed in rectangular section chassis rails. Using C section chassis rails will cause a big reduction in stiffness compared to a chassis made of rectangular section.
To summarise the ladder frame versus space frame issue a space frame equipped car has a small but significant advantage over a ladder frame equipped car assuming the chassis are equally well designed and made. It is not true that all space frames are good nor is it true that all ladder frames are much worse, weight for weight, than all space frames. There is a very big difference between good and bad space frames. This allows well designed ladder frames to equal or better many space frames.
The most common mistakes for space frames are absence of sufficient triangulation or panelling around the front suspension region and the engine bay. Poor triangulation or panelling of the rear suspension region and engine bay is common on mid engined cars. Poor triangulation or panelling of the transmission tunnel is common on front engined cars. It is important to note that for a panel to be structural it should be a welded in steel panel. Panels should be stitch welded or, preferably, continuous welded by stitching twice, the second time to fill in the gaps left the first time.
3.0The Lowcost Chassis
The chassis modifications necessary to make big improvements in many space frames are simple and do not always result in increased weight or complication. The following is a summary of my analysis of the Seven type of chassis with reference to Ron Champion’s book “build your own sportscar for as little as £250” which contains plans for a chassis of this type. Note that these figures only take into account the basic welded steel structure of tubes and panels and are subject to the usual differences between builds and the inaccuracies inherent in the simple analysis used. Alloy panels and other bolted, bonded or riveted on bodywork will increase the stiffness. Panels are assumed to be continuously welded in place by stitch welding twice, the second time to fill in the gaps left the first time.
Chassis by the book with 16 gauge sheet steel panels
The stiffness is 1187 ftlbs per degree of twist and the weight is 158 lbs.
With a welded on dashboard structure and considering some of the possible variations in the book the stiffness could be about 1400 ftlbs per degree of twist.
This is how to up rate the Lowcost chassis.
All tubes in the original design remain as in the book. Extra tubes are assumed to be 1 inch square with 16 gauge wall thickness. All steel panels, except for the seat belt mount reinforcements and rear suspension mount reinforcements, are 18 gauge.
Form a V joining the centre of tube LC to the ends of tube LD. This triangulates the front with the tubes running immediately behind the radiator. The reduction in airflow will be minimal. If radiator, fan or water pipe clearance is a problem then a diagonal or X brace across the chassis in this position may be used. Alternatively a similar modification connecting the ends of tubes FU1 and FU2 may be used but may cause clearance problems with the front of some engine ancillaries. A front V brace adds two tubes to the chassis.
Form two diagonal braces, one on each side of the chassis, between the tops of tubes LA and LB and the bottoms of tubes FU1 and FU2. This carries the triangulation of the chassis sides right to the front of the chassis and crosses the rectangular hole in each side of the chassis roughly defined by the top and bottom wishbone mounting points. Check that there is room for the steering rack. The braces add two tubes to the chassis.
Some Lowcost builders have reported that the floor is prone to flexing when thin gauge steel is used. Floor reinforcing tubes, running parallel to B2 and just in front of or under the front of the seats may be welded in, one on each side of the car. This adds two tubes to the chassis.
Weld in a panel across the bottom of the chassis between tubes E and LD. The book gives this as optional. The alloy panel referred to in the book contributes little to the chassis.
The next step is to box in the transmission tunnel from tube O3 to tube P. This makes the transmission tunnel into a welded 18 gauge steel tube enclosed on the sides, top and bottom. A hole for the gearlever will be required. A hole for the handbrake will also be required unless you decide to mount the handbrake under the dashboard as on the Caterham Seven
If you intend to establish the length of the propshaft as described in the book then you will have to leave the tunnel unfinished until after the prop dimension is taken. For final assembly it should be possible to feed the prop spline onto the gearbox spline as the area beneath the gearbox, in front of tube B2, is not panelled. The propshaft may need to be fixed to the diff during final assembly as the rear propshaft flange may be inaccessible in the finished boxed in tunnel.
The ¾ inch tubes forming the frame of the transmission tunnel do nothing if this modification is done and we can therefore take an opportunity to reduce weight. Delete tubes c, d, g, h, i, j, the two rear k tubes and the tube which connects the tops of the two rear k tubes. A single arch over tube B2 may be required to give local reinforcement to support the handbrake or gearshift mechanisms hence the retention of the front k tubes and the tube that connects their tops. Check a Caterham chassis if you find it hard to believe that tubes may be removed, it has a very light structure in this region indeed. This step removes a total of nine tubes from the chassis.
The picture shows the extra tubes.
The picture shows the extra welded in panels.
We now have a good all round improvement for not much effort:
Chassis with modified front, 18 gauge sheet steel panels and boxed in tunnel with no internal ¾ inch tubes except for front hoop and floor braces.
The stiffness is 2449 ftlbs per degree of twist and the weight is 148 lbs
Note that the weight is lower, the stiffness much higher and the number of tubes is reduced by three compared to the basic chassis built to the book which proves that extra strength need not mean extra weight or complication.
Note that if a Satchel link is used to locate a live axle or Deon axle or if an independent double wishbone suspension is used then the tubes around the back of the transmission tunnel will need to be stronger than ¾ inch and should be 1 inch square as a minimum.
I am often asked if these modifications can be done individually. The answer is yes. Here is a resume of some of the results.
Book chassis
16 gauge panels
158 lbs and 1187 ftlbs/degree of twist
18 gauge panels
142 lbs and 1169 ftlbs/degree of twist
These two results show that panel thickness has a much greater effect on weight than on stiffness.
Book chassis but with transmission tunnel removed
18 gauge panels
135 lbs and 839 ftlbs/degree of twist
This result shows that the book transmission tunnel makes a significant contribution to the book chassis.
Modifications to front of chassis as described above but with standard transmission tunnel
18 gauge panels
147 lbs and 1838 ftlbs/degree of twist
All modifications as described above
18 gauge panels
148 lbs and 2449 ftlbs/degree of twist
All modifications as described above but all chassis in 18 gauge tubes and 20 gauge panels
111 lbs and 1835 ftlbs/degree of twist
This one may be of interest to builders of bike engined cars. It is still 50% stiffer than the book chassis but is much lighter. The chassis is possibly getting a bit flimsy in some places for anything other than a very light weight car and some areas, such as the tubes carrying drive train component mounts and suspension link mounts may be better off in one inch wide by one and a half inches high 18 gauge tube or in standard 16 gauge one inch square. It is up to you to determine if your car and your welding skills are suitable for this option as stiffness is not the same as strength and the strength is less than for the book chassis. This option is certainly not advisable for car engines as they will probably require stronger tubes to properly support them. This is also not advisable if your ability to weld thin metal is not good.
4.0Larger engines in the Lowcost chassis
The standard book chassis was originally designed for small Ford engines. The book refers to engine sizes of 1100cc and 1300cc with the occasional reference to 1600cc. My high stiffness modifications or another equivalent set of modifications to improve stiffness should bring improvements for all engine sizes. For larger engines than the book specification further changes are advisable.
For slightly larger engines than the book design, about 1.6 to 2.0 litres, tubes TR1 and TR2 become unsatisfactory. This is more to do with these tubes being long and thin, and therefore tending to bend under load, than their actual size. Other tubes also benefit from changes. I would suggest increasing the sizes of some of the tubes as follows.
TR1 and 2
14gauge 1inch diameter or 16gauge 1 inch square minimum
TR3, 4, 5 and 6
16gauge 1 inch square
C, G1, G2 and E
16gauge 1 x 1.5 inches (one and a half inches deep) It may be easier to connect tube G1 and G2 to the ends of tube E
R, J1, J2, N1 and N2
14gauge 1 inch square or 16gauge 1 x 1.5 inches (one and a half inches deep)
For bigger engines than 2.0 litres further modifications are required. I suggest the following.
The book chassis stiffness is becoming marginal at this performance level so use my high stiffness modifications or another equivalent set of modifications to improve stiffness.
Replace TR1 and TR2 with a new arrangement as follows.
Add vertical tubes from the engine mount positions on tubes F1 and F2 to tubes J1 and J2. Add two new diagonals on each side of the engine bay from the bottom of the vertical tubes to the tops of FU1 and 2 and the tops of tubes H at the ends of tube Q.
Replace the engine mount plates with tubes connecting F1 to G1 and F2 to G2.
Add tubes from the inner ends of the F to G tubes to the top of the new vertical, F to J, tubes. These tubes are to support the engine mounts.
Increase the sizes of some of the tubes as follows.
TR3, 4, 5 and 6
16gauge 1 inch square
C, G1, G2 and E
16gauge 1 x 1.5 inches (one and a half inches deep) It may be easier to connect tube G1 and G2 to the ends of tube E
R, J1, J2, N1 and N2
14gauge 1 inch square or 16gauge 1 x 1.5 inches (one and a half inches deep)
Change the size of K1 and 2 to 14gauge 1 inch square or 16gauge 1 x 1.5 inches (one and a half inches deep)